Hugging Face's new "HF Skills" lets you fine-tune any open-source LLM (0.5B–70B) with just one natural language line. It auto-handles dataset checks, GPU booking, training (SFT/DPO/GRPO), monitoring, and pushing the model to the Hub — all for as little as $0.30 per run.
1 comment
[ 5.6 ms ] story [ 11.6 ms ] thread